which year the news of their rising reached Augustus in Gaul, and diverted him (see above, § 3) from his intended expedition to Britain. He was commanding in person against them in B.C. 25, but fell ill and was detained at Tarraco for some months. In the meantime the war was concluded by C. Antistius and T. Carisius, his 'legati.' Augustus himself returned home in B.C. 24. In the same year they rose again (Dion 53. 28) and seized by stratagem and killed some Roman soldiers, but were again put down by L. Aemilius.

The expressions of Od. 2. 6. 2 Cantabrum indoctum iuga ferre nostra,' and 2. 11. I 'Quid bellicosus Cantaber ... cogitet,' would be intelligible at any time during this period, and as each conquest would be thought final till the next rising, there is nothing even in the words of 3. 8. 21 (Servit Hispanae vetus hostis orae Cantaber sera domitus catena') to fix them necessarily to a single date. Other considerations perhaps place the Ode, as we shall see, either in B.C. 29 or in 25.

The final subjugation of the Cantabri by Agrippa in B.C. 19 (Dion 54. 11) does not come within the period of Odes i-iii, but is recorded in Epp. 1. 12. 26, and alluded to in Od. 4. 14. 41.

§ 7. Daci, Getae, Scythae, Geloni.-There is much vagueness in the use of these names by Horace, as indeed there is confessedly in their use by prose writers of much later date. The name 'Scythae' is the most comprehensive, being used apparently for all the tribes north of the Danube and Euxine. At times it is brought into close relation with that of the Getae (as in Od. 3. 24., 11), who again are closely connected by all writers with the Daci. At other times it is associated with the Geloni and the Tanais (Scythicus amnis,' Od. 3. 4. 36), and denotes tribes far enough to the East to interfere in Parthian politics. The names are often used merely as poetical expressions of distance, the extreme North (as in Od. 2. 20), or generally for the northern tribes, as the supposed representatives of the manlier virtues. (as in Od. 3. 24) or as the objects of the vague fears of Roman statesmen (Od. 2. 11. 1).

The Daci are mentioned by Dion 51. 22 as offering their services to Octavianus, and when their terms were declined by him joining Antony, to whom, however, they rendered little assistance, as they were quarrelling amongst themselves (see Od. 3. 6. 13).

In B. C. 30, M. Crassus, at the bidding of Octavianus, marched northward from Macedonia, and won some victories over the Daci and Bastarnae as well as the Moesi, for which in B.C. 27 he was allowed a triumph, Dion 51. 23.

From the Epitome of Livy (B. 135) it appears that Crassus was again fighting in Thrace in B. C. 25.

Florus (4. 12, § 18) speaks of Lentulus driving the Daci beyond the Danube, but no date is given. His words may be worth quoting for his mention of Cotiso (Od. 3. 8. 18) and for the illustration of Horace's expression intra praescriptum equitare,' Od. 2. 9. 23: 'Daci montibus inhaerent; Cotisonis1 regis imperio quotiens concretus gelu Danuvius iunxerat ripas decurrere solebant et vicina populari. Visum est Caesari Augusto gentem aditu difficillimam submovere. Misso igitur Lentulo ultra ulteriorem repulit ripam: citra praesidia constituit, sic tunc Dacia non victa sed submota atque dilata est. Sarmatae patentibus campis inequitant; et hos per eundem Lentulum prohibere Danuvio satis fuit.

It is obvious that there is nothing here to fix the date of the debated Ode 3. 8. The victory of Crassus will satisfy the expressions of v. 18, and so would the victory of Lentulus, but this last is itself undated.

The Eastern Scythians are named by Justinus as interfering in the quarrel between Phraates and Tiridates in Parthia (see the next section). An embassy of Scythians is said by Orosius (see Introd. to Od. 2. 11) to have come to Augustus while he was at Tarraco in Spain B.C. 25.

§8. Parthia. The defeats of Roman armies under Crassus, Decidius Saxa (the legatus of M. Antony), and M. Antony himself, in B.C. 53, 40, and 36, though the objects of frequent reference in Horace's poems, and grounds of the keen interest taken in Parthian affairs, and of the stress laid on the mission of Augustus to restore Roman prestige in the East, yet all fall without the period assigned for the composition of the Odes. The only contemporaneous event of Parthian history is that which is related by Dion 51. 18, and by Justinus 42. 5. 5. Phraates IV, to whom Orodes I had resigned his throne in B.C. 38, after some years of tyranny, provoked his subjects to the point of rebellion. He was expelled, and Tiridates, another member of the Arsacid house, though his exact relationship to Phraates is unknown, was put on the throne in his place. After a short time Phraates was restored (Justinus adds, by the intervention of the Scythians), and Tiridates fled to seek the protection of Augustus, carrying with him

1 Suetonius, Oct. 63, calls him 'Getarum rex,' and gives a story, on Antony's authority, of Augustus having at

one time promised Julia in marriage to him, and asked a daughter of his in return.

the infant son of Phraates. These events are undoubtedly the objects of reference in Od. 1. 26. 5, 2. 2. 17, 3. 8. 19, and very probably also in 1. 34. 14-16 and 3. 29. 28. If we could date them therefore with certainty we should know the earliest time at which the first-named Odes at least could have been written. And it so happens that this would incidentally throw light on one or two more points of Horatian chronology, for 3. 8 is written on an anniversary (it seems almost necessarily the first anniversary) of Horace's escape from the falling tree. To fix, therefore, the earliest date of this Ode would determine as much for the other Odes which refer to the accident, i. e. 2. 13 and 3. 4. Horace's escape again is connected (2. 17. 21-30) with Maecenas' reception in the theatre on his recovery from illness, and this in its turn gives a date of some kind to 1. 20. The date, however, on which so much depends is not itself quite free from doubt. Justinus says that Tiridates fled to Augustus, 'who was at that time fighting in Spain,' which would fix the date between B. C. 27 and 24. Dion, on the contrary, narrates the event under the year 30, and makes Tiridates find Augustus in Syria, on his progress through Asia after the battle of Actium. Mommsen (Res gestae divi Augusti, vi. 1-3) thinks the two acounts should be both accepted, as giving two stages in the negotiations of Tiridates with Augustus, but in that case the reference of Horace would probably be to the earlier one. Another point of some interest has been supposed to be involved in the date of Tiridates' flight. Two of the Odes which refer to this event (3. 8 and 29) speak also of Maecenas as burdened with cares of State in a way in which no other Ode speaks of him. Mitte civiles super urbe curas,' Tu civitatem quis deceat status Curas, et urbi sollicitus times.' These expressions have been usually interpreted of the powers which Augustus is known to have delegated to Maecenas during his own absence from Rome in the last year of the civil war. Dion 51. 3, Tac. Ann. 6. 11 Augustus bellis civilibus Cilnium Maecenatem equestris ordinis cunctis apud Romam atque Italiam praeposuit.' If the later date of these Odes were adopted it would seem necessary to assume, what is probable enough in itself, but not otherwise ascertained, that the same powers were entrusted to Maecenas during Augustus' absence in Gaul and Spain in the years B.C. 27-24.

II-THE ORDER OF THE ODES, AND THEIR DIVISION INTO THREE BOOKS.

§ 9. In the preceding pages I have assumed the correctness of the traditional view that Books i-iii form a unity. How far particular Odes or groups of Odes may have been shown to friends or given wider publicity before the whole collection was complete, it is of course beyond our power to guess: but we mean that the three Books were arranged as we have them by Horace himself, at one time, and intended to be read as a whole.

6

§ 10. This can hardly be said to rest on external evidence, for although the words of Suetonius, tribus carminum libris ex longo intervallo quartum addere,' lend themselves to the idea of such a substantial unity of the earlier Books, they do not necessitate it; but the internal evidence in its favour is very strong. In other cases where Horace's poems are divided into 'Books,' there is a corresponding difference of personal and historical background which explains the division: these three 'Books' reflect the same time; all their references to persons and events point, as we have seen, to a single period of about seven years. Whatever be their principle of arrangement, whether within the separate Books or as between them, it is not chronological'. The only Ode than can be with confidence dated as early as B. C. 30 is 1. 37. Of two Odes which can be definitely placed in B. C. 24 one is in the third Book, the other in the first (3. 14, 1. 24). Od. 1. 12 belongs almost certainly to the later years of that period. The Odes which turn on the two synchronous personal events, the poet's escape from the falling tree and the recovery of Maecenas from dangerous illness, are distributed between the three Books, and the one which must be latest in composition (1. 20) is in the first Book. The view based on such considerations is seen to harmonize with indications of unity within the poems themselves. It seems to be implied in the analogy which we notice between the

1 It will be noticed that if this is the case, and if also the existing arrangement of the Odes is the original one, it follows necessarily that the three Books must have been published simultaneously. The division of the Books and the order of the Odes as they are have the right of undisputed tradition. One MS. only departs from them (viz.

B), and that, as will be seen from the account given of it on pp. 4, 5, bears, though in another way, witness to the usual arrangement. Diomedes, the writer on metre (quoted by Priscian and therefore not later than the fifth century), refers to the Odes by their present numbering.

dedication of Epp. 1. 1 and 19, and of Od. 1. 1 and 3. 29 to Maecenas, while in each case the last poem in the collection, Epp. 1. 20, Od. 3. 3o, is reserved for the poet's literary self-consciousness. The references in Book iv. treat the three preceding Books as a whole-as e. g. in the relation of 4. I to 1. 19 and 3. 26, as one of the earliest and the latest of the love Odes of his earlier poetry. Above all, it is only when this unity is recognized that we perceive that full significance in the arrangement of the Odes which the example of the Fourth Book prepares us to expect1.

§ 11. Some kind of conscious arrangement subsequent to composition, and not chronological, is obvious on the face of the Epodes, Satires, and Epistles. There is at least the choice of the opening poem, not usually 2, if ever, earliest in date of composition, an apology for the style of writing as in Sat. ii. or a quasi-dedicatory address as in the Epodes, Sat. i, and both Books of the Epistles. And there are reasons frequently to be detected for the juxtaposition or separation of particular poems. In the Epodes, for instance, we notice especially the metrical arrangement of 1-10, 11-16, 17, and the distribution at intervals of poems upon the same subject, such as those upon Canidia (5 and 17) and those in which he touches politics (1, 7, 9, 16).

4

§ 12. It is natural that the art expended on the arrangement of the Odes should be greater. The moods which they reflect are more various, and the poems are of a kind which has more to gain by contrast or preparation. But there is a greater reason in the delicacy of the ground upon which they enter, in respect to the politics of the time. Horace has in them to justify his change of sides, neither to ignore nor to make too much of it-he has to praise with tact one, I cui male si palpere, recalcitrat '-he has with dignity and without offence to do justice to old friendships and old ideals. In this task he takes refuge in the irony, partly a method of his art, partly a natural

It may be added that if, as is generally believed, Epp. 1. 13 refers to the presentation of the Odes to Augustus, it is in accordance with the current view that Horace speaks of the poems sent as 'libelli,'' fasciculus librorum,' in contrast with his use of the singular 'libellus' of the First Book of the Satires in Sat. 1. 10. 92.

2 Epod. 1, if it refers, as seems probable, to Actium, is one of the two latest in the Book. In Epp. ii the first of the two Epistles is to be dated six or seven years later than the second.

In Sat. ii the first Satire seems to carry in it a reference to B. C. 30, the year of the publication of the Book.

3 For instances in the Satires and Epistles, see vol. ii, pp. 14 and 210.

The case of Odes iv is dealt with fully in the introduction to that Book. The art of the arrangement is more generally recognized in it than in the three earlier Books, only because the material to be disposed is smaller in quantity, and because the purpose of the Book is more fully avowed.